Temi di Discussione(Working Papers)

Industry dynamics and competition from low-wage countries:evidence on Italy

by Stefano Federico

Num

ber 879S

epte

mb

er 2

012

Temi di discussione(Working papers)

Industry dynamics and competition from low-wage countries:evidence on Italy

by Stefano Federico

Number 879 - September 2012

The purpose of the Temi di discussione series is to promote the circulation of workingpapers prepared within the Bank of Italy or presented in Bank seminars by outside economists with the aim of stimulating comments and suggestions.

The views expressed in the articles are those of the authors and do not involve the responsibility of the Bank.

Editorial Board: Silvia Magri, Massimo Sbracia, Luisa Carpinelli, Emanuela Ciapanna, Francesco D’Amuri, Alessandro Notarpietro, Pietro Rizza, Concetta Rondinelli, Tiziano Ropele, Andrea Silvestrini, Giordano Zevi.Editorial Assistants: Roberto Marano, Nicoletta Olivanti.

ISSN 1594-7939 (print)

Printed by the Printing and Publishing Division of the Bank of Italy

INDUSTRY DYNAMICS AND COMPETITION FROM LOW-WAGE COUNTRIES: EVIDENCE ON ITALY

by Stefano Federico*

Abstract

This paper analyses the effect of competition from low-wage countries on domestic activity, using data on 230 Italian manufacturing sectors between 1995 and 2007. It finds that low-wage import penetration is negatively related to employment and other measures of activity. The effect is significantly smaller in more skill, capital and R&D-intensive sectors and in more vertically differentiated sectors. There is also evidence of significant effects of low-wage competition through inter-industry linkages: employment is negatively related to low-wage import penetration in downstream sectors but positively related to low-wage import penetration in upstream sectors.

JEL Classification: F16, F14, L60, D57. Keywords: import penetration, low-wage country competition, factor proportions, inter-industry linkages.

Contents

1. Introduction.......................................................................................................................... 5

2. Methodology and data ......................................................................................................... 8

3. Main empirical findings..................................................................................................... 18

4. Heterogeneous effects across sectors................................................................................. 25

5. Inter-industry linkages ....................................................................................................... 29

6. Concluding remarks........................................................................................................... 35

Appendix ................................................................................................................................ 38

References .............................................................................................................................. 41

Tables and figures................................................................................................................... 44

_______________________________________

* Bank of Italy, Economic Research and International Relations.

1 Introduction1

The integration of low-wage countries into the world trading system has

been among the most important economic developments of recent decades.

It has involved a massive increase in the number of workers producing for

international markets, with far-reaching implications for the rest of the world,

including advanced countries. Low-wage countries’ market share of world

exports of goods has correspondingly expanded, with China’s share, for

instance, rising from less than 3 percent in 1995 to 9 percent in 2008 (United

Nations, 2009). These developments have led many observers to identify

low-wage competition as a major threat to employment and wages in the

manufacturing sector in advanced countries.

The employment and wage effects of import penetration have long been

studied. The initial evidence, pointing to a weak impact of trade on domestic

developments (Grossman, 1982; Krugman and Lawrence, 1993), has been

reconsidered afterwards (Freeman and Katz, 1991; Revenga, 1992; Sachs

and Shatz, 1994; Krugman, 2008; Bernard et al., 2002 2006; Baldwin and

Lileeva, 2008; White, 2008; Khandelwal, 2010). This was due in part to

improved methodologies and in part to significant changes in trade patterns:

not only has trade integration increased markedly since the mid-1990s, but

the differences across trading countries in terms of factor abundance have

become larger. Specifically, the set of developing countries contributing to

1I am grateful to two anonymous referees, Matteo Bugamelli, Gilles Duranton, RobertoTedeschi and seminar participants at the Rimini Conference in Economics and Finance,the EPI Conference in Parma, the European Trade Study Group in Lausanne and theDemography of Firms and Industries Conference at Paris Est Creteil University. Theviews expressed in this paper are those of the author and do not necessarily reflect thoseof the Bank of Italy.

5

a significant fraction of the increase in trade has lower wages (relative to

wages in advanced countries) than the set of developing countries that was

studied in the literature in the 1980s and early 1990s (Krugman 2008). Trade

between countries with markedly different factor intensities may have in fact

a stronger impact on reallocation across sectors than trade between similar

countries, as is suggested, for instance, by the Heckscher-Ohlin model.

This paper analyses the effects of competition from low-wage countries

on employment and other measures of sector activity in Italy. Using a panel

of 230 manufacturing sectors over the years 1995-2007, it employs variation

in import penetration across industries and over time in order to estimate

its effect on employment. It also uses instrumental variable (IV) methods

to control for the endogeneity of imports, which are clearly influenced by

demand and supply conditions in the domestic country. Specifically, this

work tackles the following questions. Does employment decrease in sectors

more exposed to low-wage import penetration, and, if so, by how much?

Does employment decrease less in capital or skill-intensive sectors, as one

would expect from factor proportions theory? Does import penetration also

have an impact on employment in upstream and downstream sectors?

The contribution of this paper to the literature is twofold. First, it extends

the analysis of the employment effects of low-wage country competition to a

new country, supplementing the available evidence, which is largely restricted

to the United States. With its relatively large and open manufacturing

industry, Italy provides a suitable setting for such an analysis. Italy’s

specialization in low-technology manufacturing sectors makes it particularly

exposed to competition from low-wage countries. Several studies using

6

data for Italy have already found evidence of significant effects of Chinese

competition on sectoral productivity (Bugamelli and Rosolia, 2006) and on

firm-level prices (Bugamelli et al., 2010).

Second, this paper explores the effects of low-wage country competition

through inter-industry linkages. The usual approach typically relates

employment outcomes in one sector to import penetration in the same sector,

thus ruling out the possibility that sector-specific shocks spill over to other

sectors. However, as suggested by Long and Plosser (1983) and Shea (2002),

supply or demand shocks to a given sector can be propagated to other sectors

through input-output linkages. We therefore innovate by linking employment

outcomes in one sector to import penetration in upstream and downstream

sectors, finding that this transmission channel is far from negligible.

We find that an increase in import penetration from low-wage countries

is associated with a decrease in employment, output, the wage bill and

the number of firms. The decrease in industry activity is smaller in more

skill, capital, R&D-intensive and vertically differentiated sectors. We also

find evidence of significant inter-industry linkages: import penetration in

downstream sectors is negatively related to employment, as it lowers the

demand for domestically produced inputs, while import penetration in

upstream sectors is positively related to employment, presumably thanks

to the benefits from using cheaper inputs.

The rest of the paper is structured as follows. Section 2 provides a

description of the theoretical framework, the estimation model, identification

issues and data. Section 3 reports the main empirical results, while Sections 4

and 5 focus on heterogeneous effects across sectors and inter-industry linkages

7

respectively. Section 6 concludes.

2 Methodology and data

2.1 Theoretical framework

This section outlines a simple theoretical model, based on Pindyck and

Rotemberg (1987), which will guide our empirical analysis. For a competitive

domestic industry, supply and demand are given by the following equations:

St = s0 + s1at + s2Pt + εst (1)

Dt = d0 + d1bt + d2Pt + εdt (2)

where Pt is price, at is a supply shift parameter (technological progress,

labour costs, etc.), bt is a demand shifter (income level, tastes, etc.) and εst

and εdt are unobserved supply and demand shocks, respectively. Equilibrium

is given by the following equation, which relates domestic demand in a given

sector to the sum of domestic supply and imports:

Dt = St + (MLWt +MOT

t ) (3)

where MLWt represents imports from low-wage countries and MOT

t imports

from other countries. The import supply curve is assumed to be upward

sloping, depending on Pt and a shift parameter ct (which reflects foreign

supply and demand conditions):

8

Mkt = mk

0 +mk1c

kt +mk

2Pt + εkmt ∀k = LW,OT (4)

It is then assumed that Yt (a measure of activity in the domestic industry:

employment, wage bill, output, profits, etc.) is a function of domestic supply

and of the shift parameter at:

Yt = i0 + i1at + i2St + εit (5)

Using (1), (2), (3) and (5), Yt can be written as a function of demand

and supply shift variables, imports from low-wage countries, imports from

the rest of the world and an error term:

Yt = ψ + αat + βbt + δMLWt + γMOT

t + εt (6)

where εt is a combination of the error terms in the supply and demand

equations (εst and εdt) and of εit. This extremely simple model states that

domestic employment in a given sector will be negatively affected by outward

shifts of the foreign supply curve.2

Further insights can be derived from the Heckscher-Ohlin model with

two additional important assumptions. First, we assume that countries

have different factor endowments: labour is relatively more abundant in

low-wage countries, capital (including human capital) is relatively more

abundant in advanced economies. Second, we assume that sectors have

2Autor, Dorn and Hanson (2011) present a simple model based on Eaton and Kortum(2002) where a reduction in trade costs or an exogenous increase in productivity in Chinaincreases the toughness of competition and reduces the demand for goods produced by aregion, thus reducing the region’s output.

9

different factor intensities: machinery and transport equipment are usually

more capital intensive, textiles and apparel are usually more labour-intensive.

The model then predicts that each country exports the good that makes

intensive use of the input that is abundant in that country (factor proportions

theory). Thus, capital-abundant countries export capital-intensive goods

while labour-abundant countries export labour-intensive goods.

This framework can also be used to explore the implications of a decrease

in the price of the labour-intensive good. This could derive, for instance,

from the removal of trade barriers in that sector, from an increase in the

overall output capacity of low-wage countries, etc. The decrease in the

price of the labour-intensive product lowers the price of labour relative

to capital. This determines an expansion of the labour-intensive sector

in the labour-abundant country, while the opposite takes place in the

capital-abundant country, with a reallocation of output and employment

away from the labour-intensive sector towards the capital-intensive sector.

For some factor endowments, the capital-abundant country will even

completely cease producing the labour-intensive good, thus moving to a

situation of complete specialization in the capital-intensive good.3

2.2 Estimation model

The estimated specification is the following:

lnLjt = αj + βt + δLWPENjt + γOTPENjt + εt (7)

3Using an assignment model with a continuum of goods and workers, Costinot and Vo-gel (2010) also show that, when the skill-abundant country opens to trade, the employmentshare in skill-intensive tasks will increase.

10

where Ljt is the level of employment in industry j at time t, αj is an

industry fixed effect, βt is a time fixed effect, LWPENjt and OTPENjt

measure low-wage countries’ import penetration and other countries’ import

penetration, respectively. A similar specification is used by Khandelwal

(2010) and, with firm-level data, by Bernard et al. (2006). The estimating

equation is a transformation of equation (6) to a panel of sectors, where the

measure of activity Yt is given by the employment level and imports are scaled

by apparent consumption. Comparing equation (7) with equation (6), one

may notice that the supply and demand shifters, at and bt, are assumed to be

captured by the set of industry and time fixed effects. These controls might

not be sufficient if there is skill-biased technological change, which causes

advanced countries to move away from unskilled industries independently

of import competition. In this case low-wage import penetration would

simply be filling a gap in unskilled production determined by a long-term

technological trend. To control for skill-biased technological change, we will

include the interaction between two-digit sector and year dummies in all

the main specifications. The effect of import penetration is then identified

using only on the basis of variation among four-digit sectors within a given

two-digit sector.

The import penetration variables require detailed discussion. First, there

is a general question of how to measure foreign competition properly, whether

using import prices (Grossman, 1982; Revenga, 1992) or import quantities

(values or volumes of imports, normally scaled on domestic consumption or

output: Freeman and Katz, 1991; Bernard et al., 2006; Khandelwal, 2010).

The use of import prices is problematic for a number of reasons. First,

11

it is well known that import prices derived from unit values suffer from a

composition bias, as the bundle of products imported in a given industry

may change over time. Second, import prices derived from direct surveys are

not available for some countries (including Italy), and even when they are

available they are much less disaggregated by product or partner than import

quantities; this is particularly important for our purposes, as this work looks

at differences between low-wage countries and other countries. Finally, if

products are differentiated an increase in the variety of imports would not

be reflected in conventional price indices, while it would correctly show up

in import quantities (Feenstra, 1994, 1995 pp. 1587-88).

These reasons lead us to build our measure of import penetration as

the share of imports in domestic consumption of the goods produced by an

industry (where domestic consumption is the sum of imports and domestic

production minus exports), as in Bernard et al. (2006) and Khandelwal

(2010):

LWPENjt =MLW

jt

M j +Qj −Xj

(8)

OTPENjt =MOT

jt

M j +Qj −Xj

(9)

where MLWjt is imports from low-wage countries in Italy’s sector j at time

t, MOTjt is imports from other countries (imports from all countries minus

imports from low-wage countries), M j is sector j’s imports from all countries,

Qj is sector j’s domestic production andXj is sector j’s exports. All variables

are in value terms. As in Auer and Fischer (2010), the denominator (imports

12

from all countries plus domestic output minus exports) is averaged over

1995-2006, in order to reduce the correlation between domestic production

(included in the denominator of (8) and (9)) and domestic employment

(dependent variable in (7)). Our measure of low-wage import penetration

equals 0.10 if imports from low-wage countries account for 10 percent of

average domestic consumption in the sector in question.

A country is classified as low-wage if its GDP per capita in current U.S.

dollars is less than 10 percent of U.S. GDP per capita in 2006. As in the

previous literature, we use non-PPP-adjusted data, which are available for

a wider number of countries than PPP-adjusted data. Our threshold is in

between the 5 percent used by Bernard et al. (2006) and Khandelwal (2010)

and the 20 percent threshold used by Auer and Fischer (2010).4 We choose a

fixed cut-off, based on GDP per capita ratios in the final year of our sample,

rather than a time-varying cut-off, in order to avoid noisy fluctuations due

to countries crossing the threshold in either direction over the years. We also

believe that since our measure is based on imports from all countries below

a given threshold, it is preferable to one based on imports from just one

country, such as China. The latter would be biased, for instance, if imports

from China are only replacing imports from other low-wage countries. The

list of low-wage countries is presented in the Appendix.

A final remark on the import penetration variables concerns their

dynamics. Domestic employment is likely to react to foreign competition

with a certain lag. We will therefore estimate employment at time t as a

4Since Italy’s GDP per capita was 72 percent of U.S. GDP per capita in 2006, thelow-wage cutoff corresponds to 14 percent (= 10/72 percent) of Italy’s GDP per capita.We will control the sensitivity of our results to alternative cutoffs.

13

function of import penetration at time t − 2 (as in Bugamelli and Rosolia,

2006). The results are robust to different lag structures.

2.3 Identification

Equation (7) can be estimated consistently with ordinary least squares,

provided that the explanatory variables are not correlated with εt (which

is a combination of εst and εdt). There are several reasons why this condition

may not hold true in the case of import penetration variables. For instance,

an unobserved positive demand shock will raise prices and therefore will also

raise imports, determining a correlation between εt and import penetration.

A similar line of reasoning holds for unobserved supply shocks, such as an

increase in labour costs or a decrease in productivity. As Pindyck and

Rotemberg (1987) point out, this issue becomes even more problematic when

imports are highly price-elastic, because in this case they will respond more

strongly to domestic shocks having an impact on prices. Ideally, one needs

an instrument which is correlated with imports but not with unobserved

domestic supply and demand shocks.

As argued by Auer and Fischer (2010), although integration of low-wage

countries into the world trading system has been relatively speedy, its

cumulative impact cannot be evaluated with an event-study approach based

on trade liberalizations or sudden tariff reductions. Each of the various IV

methods suggested in previous studies has its own strengths and weaknesses,

and since the exclusion restrictions are usually not testable, it is important to

compare results across different IV methods. We therefore report the results

14

obtained with three alternative IV approaches.

As our first set of instruments, we compute LWPEN and OTPEN using

four-digit industry data for the United States (LWPENUS and OTPENUS).

These instruments, which are strongly correlated with the instrumented

variables, are unlikely to be correlated with supply shocks in Italy, given the

large differences between the two countries in terms of labour markets, legal

institutions, monetary policy, etc. They are also unlikely to be correlated

with demand shocks: Italy accounts for a very small fraction of U.S. imports

(around 3 percent); the U.S. takes only 7 percent of Italy’s exports and

accounts for less than 3 percent of Italy’s total sales by manufacturing firms.5

A similar approach, where imports from China to the U.S. are instrumented

with imports from China to other high-income countries, is used by Autor et

al. (2011).

The second set of instruments draws on Auer and Fischer (2010), who

start by noticing that, according to factor proportions theory, low-wage

countries should increase their exports more in labour-intensive or low-skill

intensive sectors as their output capacity grows. They therefore suggest

interacting industrial production in low-wage countries (which measures

their overall output capacity) with an industry-specific indicator of labour

intensity. This gives rise to a time and industry-varying instrument, where

the time variation derives from the industrial production index and the

industry variation derives from the labour intensity indicator. In a similar

way, we collect data on industrial production for the main low-wage countries

5Italy’s market share in U.S. imports tends to be small for the vast majority of sectors.Excluding sectors with a market share larger than 10 or 20 percent of U.S. imports doesnot have any effect on our results.

15

in our dataset (China, India and Indonesia, using the 10 percent GDP per

capita threshold) and for a group of advanced economies. Industry-level

measures of labour intensity computed on data for Italy or other countries

might be endogenously influenced by import penetration. We therefore prefer

to interact low-wage industrial production with an industry-level indicator

of routine or manual tasks intensity (IPLW ∗ ROUTINE), derived from

O*Net database (Acemoglu and Autor, 2010). We also instrument import

penetration from the rest of the world, by interacting advanced economies’

industrial production with an industry-level indicator of non-routine or

creative tasks intensity (IPOT ∗ NONROUTINE), also from the O*Net

database. In other words, we expect that a given increase in industrial

production in low-wage countries will increase their exports to Italy relatively

more in sectors that are more intensive in routine or manual tasks.

Conversely, we expect that a given increase in industrial production in

high-wage countries will increase their exports to Italy relatively more in

sectors that are more intensive in non-routine or creative tasks.

The third and final approach is based on Bernard et al. (2006) and

Khandelwal (2010), who instrument import penetration with industry-year

measures of transport costs, tariffs and exchange rates. The idea behind

this approach is that these variables shift the import supply curve, while

being largely exogenous to demand or supply shocks in the domestic

industry. Exchange rate movements mainly reflect macroeconomic conditions

in the exporting countries. As in Revenga (1992), we compute an

industry-weighted geometric average of nominal exchange rates (EXCHLW ,

EXCHOT ). Transport costs (TRANSPORTLW , TRANSPORTOT ) and

16

tariffs (TARIFFLW , TARIFFOT ) are computed using data on duties and

freight costs for U.S. imports, in order to reduce endogeneity concerns (for

instance, tariffs may be raised in order to protect domestic industries from

large increases in import penetration). All variables are separately computed

for low-wage countries and for the rest of the world.

2.4 Data description

Our data come from two main sources: Structural Business Statistics (SBS)

and External Trade Statistics (COMEXT). Both sources are based on

information supplied by EU Member States’ national statistical institutes

(Istat, for Italy) and publicly released by Eurostat, the Statistical Office of

the European Union.

From SBS we take data on employment and other structural variables

for each Italian four-digit manufacturing sector. The industry classification

corresponds to NACE revision 1. SBS data are available for Italy on an

annual basis for the years 1995-2007. Data are collected by Istat using two

firm-level surveys: the system of economic accounts in enterprises, which

includes firms with 100 or more workers, and the sample survey on small

and medium-sized enterprises for firms with up to 99 workers. Survey data

are then integrated with administrative data (balance sheets) and statistical

techniques for treatment of non-response units and extrapolation to the

universe (Istat, 2009). There are some conceptual differences between SBS

data and national accounts data: in particular, SBS data are usually not

adjusted for informal businesses (OECD, 2006). Overall, SBS data in our

17

sample represent on average more than 90 percent of both manufacturing

employment and value added in national accounts data (Table 1).

From the COMEXT dataset we obtain annual values of Italy’s imports

and exports by partner country and product code, for the years 1988-2006.

Products are defined according to the Combined nomenclature, which is

based on the Harmonized system (HS) classification and includes about

10,000 eight-digit codes. Product-level data were then aggregated to

the four-digit Classification of products by activity (CPA) level, using

concordance tables available on the Eurostat website. Since there is a

one-to-one correspondence between CPA and NACE, trade data can be

easily linked to employment data at the four-digit level. Using data on GDP

per capita from the IMF World Economic Outlook database, one can then

distinguish between imports from low-wage countries and imports from other

countries. Table 2 reports summary statistics. Sources and definitions for all

variables are presented in the Appendix.

3 Main empirical findings

3.1 Preliminary evidence

Figure 1 plots the 1996-2007 changes in employment against the 1995-2006

changes in low-wage import penetration for 211 Italian manufacturing

sectors, using unweighted data (left panel) and employment-weighted data

(right panel). Several observations can be made. First, low-wage import

penetration increased in the vast majority of sectors. Its increase did not

18

exceed 5 percentage points in most cases, although there are larger increases

(up to 30 percentage points) in a few sectors.

Second, there is a wide variability across sectors in terms of employment

changes, even if we ignore outliers, which generally correspond to very small

sectors. Overall, 98 sectors recorded employment expansion and 113 sectors

employment contraction.

Third, there is a negative relationship between changes in low-wage

penetration and changes in employment. Simple OLS regressions of

employment changes on low-wage import penetration changes yield negative

coefficients, which are statistically significant and have a similar slope. The

R-squared equals .137 using unweighted data, but rises to .319 using weighted

data. As a first approximation, these results suggest a potentially significant

effect of low-wage import penetration on employment decline at the industry

level.

An alternative way to look at the extent of competition from low-wage

countries is the following. Sectors in which low-wage import penetration

was greater than 10 percent accounted for less than 2 percent of Italy’s

total manufacturing employment in 1995 but almost 20 percent in 2006.

This suggests that low-wage competition was not confined to a few small

industries. Italy’s specialization structure, with a relatively large share of

low-technology sectors, may explain this pattern: in 1995 the textiles, apparel

and leather sector still accounted for about 20 percent of manufacturing

employment, a much larger share than in other advanced economies.

19

3.2 OLS estimates

Table 3 reports OLS estimates of equation (7). All estimates are weighted

using average employment in each sector. The reason is that SBS data

are based on sampling techniques, which estimate large sectors with more

precision than small sectors. Moreover, this procedure attenuates the impact

on our results of extreme changes in employment in very small sectors.

Columns (1)-(2) are estimated using fixed effects (FE), while columns (3)-(4)

are estimated in first difference (FD) and columns (5)-(6) are estimated in

long difference (i.e. changes between 1996 and 2007, LD).

Starting with column (1), the coefficient on low-wage import penetration

(LWPEN) is negative (-1.766) and statistically significant at the one percent

level. The coefficient on other countries’ penetration (OTPEN) is instead

not different from zero. These results are in line with the notion that trade

has a larger effect on domestic developments when it takes place between

countries with very different factor endowments than between countries with

similar factor endowments. The fit of the regression is good, as shown by the

within R-squared equal to .143.6

Column (2) includes two-digit sector-year pair fixed effects, as in

Khandelwal (2010). This is an additional control that captures time-varying

demand or supply shocks common to all four-digit sectors within the same

two-digit sector. This specification thus exploits variation in employment

and import penetration within two-digit sectors. While its magnitude now

equals -.580, the coefficient on LWPEN is still significant at the one percent

6The coefficient on LWPEN remains almost unchanged if OTPEN is excluded fromthe regression. Notice that the two variables are only weakly correlated (.199).

20

level. OTPEN is again not different from zero. Since it controls for a large

set of supply and demand shocks, this is our preferred OLS specification.

The alternative specifications in first difference and long difference

(columns (3)-(6)) further support the findings of a negative relation between

employment and low-wage import penetration and no relation between

employment and other countries’ import penetration. Only in column (4),

where the first difference specification includes two-digit sector-year fixed

effects, does LWPEN turn out to be not significantly different from zero.

We do not see this as a serious issue, since first-difference estimates mainly

capture very short-run movements while our predictions are more likely to

hold in a longer-run framework. In fact, long-difference estimates show larger

and significant coefficients for LWPEN , while the coefficients for OTPEN

are again not different from zero.7

To evaluate the quantitative impact, a one standard deviation increase in

low-wage import penetration in a four-digit sector (which corresponds to an

increase of about 7 percentage points) is associated with a 4 percent annual

decline in the employment of the same sector, according to our preferred

specification in column (2). Our results are in line with previous evidence.

Khandelwal (2010) finds that a ten-percentage-point increase in low-wage

import penetration (which corresponds to almost one and a half standard

deviations in our sample) decreases employment by 6 percent on average.8

7In the LD specification import penetration variables are one-year lagged (1995-2006change, compared to 1996-2007 change in employment), rather than two-year lagged as inthe other specifications.

8Bernard et al. (2006) find that a one standard deviation increase in low-wage importpenetration is associated with a 2-percentage-point decrease in annual employment growthfor surviving plants and a 2.2-percentage-point increase in the probability of plant closure.Their findings are based on plant-level data and therefore cannot be easily compared to

21

These figures should not be interpreted as a measure of the overall effect of

trade on employment, since they take into account only import competition

but not the increased opportunities for exports toforeign markets or the gains

from a larger variety of imports. Complete welfare calculations are beyond

the scope of this paper but could fruitfully follow the approach recently

suggested by Arkolakis et al. (2012), which makes use of a very parsimonious

set of sufficient statistics.

3.3 IV estimates

OLS estimation of equation (7) might be biased if import penetration

variables are correlated with unobserved demand or supply shocks in the

domestic economy. We now report the results of the three IV methods

presented in Section 2.3, starting with the first-stage results (Tables 4 and 5

for LWPEN and OTPEN , respectively).

Looking at columns (1)-(2) in both tables, instruments based on import

penetration in the U.S. turn out to be very closely correlated with our

endogenous variables. Despite a lower goodness of fit, the results using the

Auer and Fischer (2010) IV method are also quite satisfactory (columns

(3)-(4)). Low-wage import penetration is positively correlated with the

interaction between low-wage countries’ industrial production and routine

tasks intensity and negatively correlated with advanced countries’ industrial

production interacted with non-routine tasks intensity (Table 4, column (3)),

while the opposite holds for import penetration from the rest of the world

(Table 5, column (3)). The correlation becomes very weak instead once we

ours.

22

include the interaction between two-digit sector and year dummies in column

(4); this reflects the fact that the variation in our measures of routine and

non-routine tasks intensity is mainly across rather than within two-digit

sectors. The third IV method (columns (5)-(6)), where we use tariffs,

transport costs and exchange rates as instruments, is the least powerful,

although the F statistics are in line with other studies using a similar

approach (Khandelwal 2010). This is explained by their low variability across

time and sectors, especially when we control for the interaction between

two-digit sector and year dummies.

Table 6 reports the second-stage results for each of the three IV methods.

Overall, the coefficient on low-wage import penetration is always negative

and significant except in the last specification (where the power of the

third set of instruments is very low after controlling for the interaction of

two-digit sector and year dummies). Its magnitude is always larger than the

OLS coefficient. This is consistent with evidence reported by Bernard et

al. (2006) and Khandelwal (2010), and might be explained by non-classical

measurement error in the endogenous variable (Kane et al., 1999). While

there is some variability in the point estimates, the results are qualitatively

similar using three very different IV approaches and are almost always robust

to the inclusion of the set of two-digit sector and year interacted dummies.

The coefficient on import penetration from the other countries is generally

not different from zero. Overall, the evidence based on various IV approaches

is consistent with the evidence based on OLS.

23

3.4 Robustness

For robustness analysis, we run specifications of equation (7) with alternative

measures of aggregate activity in a sector: value added, sales, wage bill and

number of firms (Table 7). We use our baseline OLS specification in column

(2) of Table 3, where we control for the interaction of two-digit sector and year

dummies.9 The coefficient on LWPEN is always negative and significant.

The coefficient on OTPEN is not significant, except when the dependent

variable is the number of firms. A first interesting result is that low-wage

penetration turns out to have a smaller effect on the number of firms than

on employment. This implies that the scale of firms (measured by average

employment per firm) actually does not increase in response to low-wage

competition, contrary to predictions based on Lawrence and Spiller (1983).

A second interesting result is that the fall in sales and value added derived

from low-wage import penetration is larger than the fall in employment,

as also found by Khandelwal (2010). This implies a negative effect of

low-wage import penetration on various measures of labour productivity

(although the effect is only significant for value added per employee). This

finding is not consistent with the implications of a Melitz (2003) model,

where average productivity in a sector rises after trade liberalization due to

output reallocation from less efficient to more efficient producers. A possible

explanation is related to the rigidity of labour markets and to the existence of

adjustment costs in quickly reducing the workforce after a negative demand

shock.

9We check the sensitivity of our results to IV estimates, using our first approach basedon U.S. measures of import penetration, and find that the results are very similar.

24

We also check the sensitivity of our results to two additional assumptions.

One is the lag structure with which low-wage import penetration is assumed

to influence employment levels. The other is related to the definition of

low-wage countries. The first three columns of Table 8 report the results

of our baseline specification in which both import penetration variables

have been included at either a one-year, three-year or four-year lag. The

results do not depend on the choice of a specific lag structure. The second

assumption is the definition of low-wage countries as countries whose GDP

per capita was less than 10 percent of U.S. GDP per capita in the final

year of our sample. We compute low-wage import penetration (and other

countries’ import penetration) changing the GDP per capita threshold to

either 5, 20 or 30 percent. The last three columns of Table 8 show again

that LWPEN is always negatively and significantly related to employment,

while the coefficient on OTPEN is never significantly different from zero.

4 Heterogeneous effects across sectors

4.1 Factor intensity and quality ladder

We then analyse whether the employment effects of competition from

low-wage countries are heterogeneous across sectors. Our hypothesis is that

the impact of low-wage competition should be smaller in sectors that are

intensive in inputs which are scarcer in low-wage countries. This follows

from the factor proportions theory, according to which each country produces

goods that are intensive in inputs that are relatively abundant in that

25

country: the opening to trade with low-wage countries should therefore push

advanced countries’ specialization towards sectors that make intensive use of

relatively abundant factors such as capital, skilled labour and R&D.

We also consider quality ladders, which measure the degree of vertical

differentiation in a given sector. Khandelwal (2010) shows that the

employment effects of low-wage competition are smaller in sectors with

a longer quality ladder. Its model assumes that products are not only

horizontally differentiated (as is usually assumed in monopolistic competition

models) but also vertically differentiated (low-quality vs. high-quality

goods). In long-ladder sectors, firms in the advanced countries can use their

comparative advantage factors in order to specialize atop the quality ladder,

while in short-ladder sectors the scope for competition on quality is reduced.

We modify regression (7), including the interaction between LWPEN

and capital, skill, R&D intensity and a measure of quality ladder length.

Capital intensity is proxied by gross tangible investments divided by the sum

of gross tangible investments and the wage bill. Skill intensity is measured by

the number of non-production workers divided by the number of production

and non-production workers. R&D intensity is equal to the ratio of R&D

expenditures to sales. Ladder lengths are taken from Khandelwal (2010) and

adapted to our more aggregate industry classification. To avoid endogeneity,

we fix these variables at their initial values. Further details on these variables

are presented in the Appendix.

We include the interaction variables in the baseline OLS specification from

column (2) of Table 3, where we control for the interaction between two-digit

sector and year dummies. The interaction variables are included one by one

26

in columns (1)-(4) of Table 9 and simultaneously in column (5). As expected,

their sign is positive and they are always significant (except for capital

intensity), suggesting that the employment effects of low-wage penetration

are smaller for skill and R&D-intensive sectors and for sectors with a longer

quality ladder. These results are consistent with factor proportions theory,

Italy being relatively more intensive in such inputs than low-wage countries.

They are also in line with the idea that in long-ladder markets firms are less

exposed to low-wage competition than in short-ladder markets.

The lower panel of Table 9 reports the effect of low-wage import

penetration on employment at selected values of the factor intensity variable

(the mean minus one standard deviation and the mean plus one standard

deviation). In industries with low skill, capital and R&D intensity or a

short quality ladder, the effect of low-wage import penetration is strongly

negative and significant. Conversely, in industries with high endowments of

these factors or long quality ladder, there is usually no significant effect of

competition from low-wage countries.

4.2 Quality ladder and export unit values

The quality ladder hypothesis suggests that the effect of low-wage

competition is smaller in sectors with a greater scope for vertical

differentiation. In this case firms in advanced countries are able to increase

the quality of their products in order to protect themselves from foreign

competition. We provide further evidence on this hypothesis, looking at the

relation between low-wage competition and export unit values. Export unit

27

values can be considered to some extent a good proxy of quality and have

been used as such, for instance, by Schott (2004, 2008) and Fontagne et al.

(2008).

We look at the relation between low-wage import penetration and Italy’s

export unit values, relative to low-wage countries’ unit values. This allows

us to determine whether the response to low-wage competition includes a

process of quality upgrading, with firms in the advanced country improving

the quality of their products. The quality ladder hypothesis predicts that

Italy’s export unit values, relative to low-wage countries’ unit values, should

increase in sectors that are more exposed to competition from low-wage

countries and that this increase should be concentrated in sectors with a

long quality ladder.

To this end, we use the BACI dataset provided by Cepii (Gaulier and

Zignago, 2010), which reports harmonized trade data on values and quantities

(in tons) at the HS six-digit level for a large sample of countries. We regress

relative export unit values (Italy’s unit values relative to low-wage countries’

unit values) in a given HS six-digit product on low-wage import penetration

and import penetration from other countries in the corresponding NACE

four-digit sector. We also include the interaction between import penetration

variables and quality ladder. The results are reported in Table 10. We

always control for the interaction between two-digit sector and year dummies.

We also control for four-digit sector fixed effects in columns (1)-(2) and for

six-digit product fixed effects in columns (3)-(4), in order to take into account

sector or product-specific trends in relative unit values.

The results in columns (1) and (3) show that, when we look at an

28

average-ladder industry, there is no significant relation between changes in

relative unit values and changes in competition from low-wage countries or

from the rest of the world. When we introduce the interaction between

import penetration and ladder length in columns (2) and (4), however, we

find that relative export unit values increase more in sectors that are more

exposed to low-wage country competition and that have a long quality ladder,

as shown by the positive and significant coefficient of the interaction term

between low-wage import penetration and quality ladder. No such effect is

found instead for competition from other countries, as the interaction term

is not significantly different from zero. These findings are consistent with

the quality ladder hypothesis, suggesting that quality upgrading may be a

way to respond to competition from low-wage countries, especially in sectors

with a greater scope for vertical differentiation. Evidence along similar lines

is provided by Bloom et al. (2011), who show that firms tend to innovate

more if they are exposed to competition from China.

5 Inter-industry linkages

5.1 Motivation

Thus far it has been assumed that import penetration in one sector may have

effects only on employment in the same sector. However, there are reasons

to believe that this assumption might be too restrictive. The main reason

derives from inter-industry linkages. Firms in one sector may purchase inputs

from other sectors (upstream sectors) and sell their own goods as inputs

29

to other sectors (downstream sectors). They could therefore be affected by

import penetration in upstream or downstream sectors. One could expect, for

instance, that higher import penetration in downstream sectors may reduce

the demand for domestically produced inputs used by those sectors (Pindyck

and Rotemberg, 1987). Import penetration in upstream sectors could instead

have a positive impact, if it increases firms’ efficiency through the availability

of cheaper inputs (Altomonte et al., 2008). More generally, a theoretical

justification is provided by Long and Plosser (1983) and Shea (2002) with

a multi-sector model in which sector-specific supply or demand shocks are

propagated to the other sectors by input-output linkages.

A second reason is that firms usually produce several products, which do

not necessarily belong to a single sector. Multi-product firms are classified in

SBS data according to the sector of their largest product. Thus, if 51 percent

of a firm’s production is in one sector and 49 percent in a different one, the

firm (together with all its employees) will be classified by SBS data in the

former sector, although in reality it will be affected by foreign competition in

both sectors. In this context, an analysis at a more aggregate industry level

may capture competition effects that do not appear at a more disaggregate

level for measurement issues.

5.2 Methodology

To address these issues, we first aggregate our data from the four-digit level to

the two-digit level (21 sectors) and re-estimate equation (7). We then build

the following two indicators of low-wage import penetration in upstream

30

sectors and downstream sectors:

UPLWPENjt =∑i 6=j

αijLWPENit (10)

DOWNLWPENjt =∑k 6=j

βjkLWPENkt (11)

where αij is the share of inputs purchased by sector j from sector i in total

manufacturing inputs purchased by sector j and βkj is the share of output

purchased by sector k from sector j in total manufacturing output sold by

sector j (similarly to Smarzynska Javorcik, 2004). UPLWPEN is therefore

a weighted average of low-wage import penetration in upstream sectors, while

DOWNLWPEN is a weighted average of low-wage import penetration

in downstream sectors. To control for the potential endogeneity of the

input-output matrix to changes in import penetration, we use Germany’s

input-output matrix instead of Italy’s. We also fix input-output coefficients

at their value in the year 2000 rather than using time-varying coefficients.

Input-output coefficients are strongly correlated both over time and across

countries, and our results are robust to several specifications.

5.3 Results

We start by reporting OLS results in Table 11. We always control for

two-digit sector fixed effects and year dummies. Columns (2), (4) and (6) also

include controls for the interaction between macro-sector and year dummies,

where macro-sectors are aggregations of two-digit sectors based on NACE

31

letters or groups of letters (examples of such groups of sectors are “textiles,

wearing apparel and leather” or “coke and petroleum products, chemicals and

chemical products, rubber and plastic products”). This aims at controlling

for domestic supply and demand shocks, as in the previous section with

four-digit level data.10 LWPEN has a negative and significant effect on

employment at the two-digit sector level in every specification. OTPEN

is also negative and significant in almost all specifications, although its

coefficient is once again much smaller.

Using the most conservative estimate in column (2), where we control

for the interaction between macro-sector and year dummies, a one standard

deviation increase in low-wage import penetration in a two-digit sector (which

corresponds to an increase of 3.5 percentage points) is associated with an

annual decline in the employment of the same sector of 3.6 percent. This

is very much in line with the results based on four-digit sectors (where the

implied fall in employment was estimated to be 4 percent).11

We then include our measures of low-wage import penetration in

downstream and upstream sectors (columns (3)-(4)). Employment in a

given sector is negatively associated with low-wage import penetration in

downstream sectors. This could be explained by a “competition” effect,

where import penetration in downstream sectors reduces the demand for

10Admittedly sector aggregations might be too broad to capture all relevant supply anddemand shocks.

11This could be explained by two different effects that offset each other. On one hand,two-digit level estimates could be higher than four-digit level estimates if they pick upcompetition effects that are not restricted to four-digit sectors (for instance, if there issome product substitutability across different four-level sectors). On the other hand, two-digit level estimates could be lower than four-digit level estimates if firms react to foreigncompetition by moving into other four-digit sectors within the same two-digit class.

32

domestically produced inputs. In other words, imports do not only displace

domestic producers in the same sector, but also displace domestic suppliers

that used to sell their output to that sector.

Employment is instead positively related to import penetration in

upstream sectors. A potential explanation is a “productivity” effect, where

firms are able to buy cheaper inputs from their domestic suppliers thanks

to the latter’s access to imports from low-wage countries. A similar

mechanism has been studied by a recent literature in the context of trade

liberalization and imported inputs. (Amiti and Konings, 2007, Amiti and

Davis, 2008, Topalova and Khandelwal, 2011). These studies show that trade

liberalization has an effect on firms’ activity not only throught lower output

tariffs but also through lower input tariffs. Specifically, the fall in input

tariffs seems to be associated with an increase in firm-level productivity or

wages, especially for importing firms. The intuition behind this result is that

lower input tariffs allow firms to have access to cheaper, higher quality or

more varied inputs which were not available before trade liberalization. Our

study points to an analogous process where firms are able to expand their

activity as their inputs become cheaper thanks to the increase in imports

from low-wage countries.

The results are generally robust to the inclusion of other countries’ import

penetration in upstream or downstream sectors, whose coefficients are small

or not significant (columns (5)-(6)).

Table 12 reports IV estimates, where LWPEN and OTPEN , as well

as all their upstream and downstream measures, are instrumented using

the corresponding values for the United States. We are not able to apply

33

the second approach, since we would have more endogenous variables than

instruments, while the third approach does not have a good explanatory

power in the first stage. The results are quite similar to the OLS estimates.

Overall, these results suggest that low-wage import penetration has

significant effects through inter-industry linkages: employment is negatively

related to higher penetration in downstream sectors, which reduces the

demand for domestically produced inputs, while it is positively related to

higher penetration in upstream sectors, presumably through a productivity

channel.

Finally, it might be interesting to look at a few concrete examples. Figure

2 plots the employment change (the residual from regressing the change

in employment over the years 1995-2007 on the change in LWPEN and

OTPEN) versus the change in LWPEN in downstream sectors (left panel)

or the change in LWPEN in upstream sectors (right panel). The slope

is negative in the former case but positive in the latter, consistently with

our estimates. An example of the negative relationship between employment

change and import penetration in downstream sectors is given by the textiles

sector (17), which sells one third of its output (excluding the output sold to

itself and final demand) to the apparel sector. Low-wage imports strongly

increased in the apparel sector, which explains the very large increase in

LWPEN in downstream sectors for the textiles sector. Stronger competition

from low-wage countries in the wearing apparel reduced demand for the

intermediate inputs usually provided to that sector by domestic textiles firms.

A similar story can be told for the wood and wood products sector (20): one

of its main customers is the furniture, toys and other manufacturing products

34

sector (sector 36), where low-wage imports have grown very fast.

An example of the positive relation between employment and low-wage

competition in upstream sectors is given by the fabricated metal products

sector (28), which gets inputs mainly from basic metals sector (27). A strong

increase in LWPEN in basic metals might have provided advantages, in

terms of cheaper inputs, to the fabricated metal products sector, which was

therefore able to expand significantly.

6 Concluding remarks

This paper starts by noticing the growing competition from China and

other low-wage countries, as suggested by their increasing market share of

world exports. The high level of their exports, together with their factor

endowments, heavily unbalanced in favour of labour, suggests that low-wage

competition may have significant implications on industry dynamics in

advanced economies. This paper provides an empirical analysis of this

hypothesis, using a panel of 230 Italian manufacturing sectors over the years

1995-2007.

The main findings can be summarized as follows. First, an increase in

low-wage import penetration is associated with a decrease in employment,

output, the wage bill and the number of firms. According to our most

conservative estimate, a one standard deviation increase in low-wage import

penetration in a given sector is associated with an annual decrease in

employment of 4 percent in the same sector; by contrast, import penetration

from richer countries does not seem to be significantly related to employment.

35

While the sign of this result is not particularly surprising, it was important

in our view to measure the quantitative effect of low-wage competition on

employment and on the other industry variables and to assess its sensitivity

to IV methods.

Second, the decrease in employment tends to be smaller in more skill,

capital and R&D-intensive sectors. This is consistent with factor proportions

theory, which predicts that trade leads each country to specialize in those

sectors which are intensive in factors that are relatively abundant in that

country. We also find that competition from low-wage countries has a

smaller effect on more vertically differentiated sectors, and that in those

sectors export unit values, which are a proxy for the quality of products,

have increased, in line with a quality upgrading hypothesis.

Third, there is evidence of significant inter-industry effects of import

penetration through upstream and downstream channels. Employment is

negatively related to higher penetration in downstream sectors, which reduces

the demand for domestically produced inputs, while it is positively related

to higher penetration in upstream sectors, probably reflecting productivity

gains due to cheaper inputs. This result is in line with recent studies showing

that trade liberalization does not only affect demand for firms’ output but

also firms’ access to imported inputs. It also suggests that studies looking at

the consequences of low-wage competition on specific industries or firms may

offer only a partial view of the overall developments in the economy. The

analysis of multi-sector models in an open economy could provide interesting

insights in this regard.

Our work should not be interpreted as a full assessment of costs and

36

benefits of trade, which are only partially discussed here. Such an assessment

would require considering the gains from new markets available for exports,

those from importing new varieties of goods and so forth. It would also

require a more general-equilibrium approach in which one would ideally

observe the movement of workers away from sectors exposed to low-wage

competition towards other sectors, including service sectors, or their going

through spells of unemployment or leaving the labour force. Alternatively,

the approach based on sufficient statistics proposed by Arkolakis et al. (2012)

could be followed. The point of this paper is simply to show that the effects of

trade with low-wage countries can be highly significant in specific industries,

and that the corresponding changes in the structure of the economy are

consistent with factor proportions theory.

A limitation of this work is that it does not address within-industry

changes. Firms belonging to the same sector may be differently affected,

given the marked firm heterogeneity that characterizes many sectors.

Different types of workers may also be differently affected, with low-skilled

workers typically expected to lose more than high-skilled workers from trade

with low-wage countries. Finally, the effects of import penetration may be

particularly significant in some regions, where there is a larger concentration

of sectors exposed to foreign competition (Autor et al., 2011, Kandilov, 2009).

Clearly, further research is needed on these issues.

37

Appendix

Classification of low-wage countries. A country is classified as low-wage if its GDP percapita in 2006 is less than 10 percent of U.S. GDP per capita (the cut-off corresponds to 14percent of Italy’s GDP per capita). GDP per capita is non-PPP adjusted, at current pricesand in U.S. dollars. Data are taken from the IMF World Economic Outlook database, April2009 release. List of low-wage countries: Afghanistan; Albania; Algeria; Angola; Armenia;Azerbaijan; Bangladesh; Belarus; Belize; Benin; Bhutan; Bolivia; Bosnia and Herzegov-ina; Bulgaria; Burkina Faso; Burundi; Cambodia; Cameroon; Cape Verde; Central AfricanRepublic; Chad; China; Colombia; Comoros; Congo, Democratic Republic of; Congo, Re-public of; Cote d’Ivoire; Djibouti; Dominican Republic; Ecuador; Egypt; El Salvador;Eritrea; Ethiopia; Fiji; Gambia; Georgia; Ghana; Guatemala; Guinea; Guinea-Bissau;Guyana; Haiti; Honduras; India; Indonesia; Iran; Iraq; Jamaica; Jordan; Kenya; Kiribati;Kyrgyz Republic; Lao People’s Democratic Republic; Lesotho; Liberia; Macedonia, For-mer Yugoslav Republic of; Madagascar; Malawi; Maldives; Mali; Mauritania; Moldova;Mongolia; Morocco; Mozambique; Myanmar; Namibia; Nepal; Nicaragua; Niger; Nigeria;Pakistan; Papua New Guinea; Paraguay; Peru; Philippines; Rwanda; Samoa; Senegal; Ser-bia; Sierra Leone; Solomon Islands; Sri Lanka; Sudan; Suriname; Swaziland; Syrian ArabRepublic; Sao Tome and Principe; Tajikistan; Tanzania; Thailand; Timor-Leste, Dem.Rep. of; Togo; Tonga; Tunisia; Turkmenistan; Uganda; Ukraine; Uzbekistan; Vanuatu;Vietnam; Yemen, Republic of; Zambia; Zimbabwe.

Employment and other industry variables. Italy’s employment and other industry variablesare taken from the Eurostat SBS database. Data are at the four-digit level of NACErev.1 classification and cover the years 1995-2007. Data are consistent over time, althoughsampling techniques of the surveys changed in 1998. In addition, since 1999 administrativedata on balance sheets and the number of employees have also been used. In 2005, thecoverage rate was 49.6 percent for the survey of firms with at least 100 workers (61.4percent in terms of employees and 66.2 percent in terms of value added) and 3.9 percentfor the survey of firms with up to 99 workers (Istat, 2009). The wage bill, value addedand sales are deflated by producer prices (also taken from Eurostat). CAP is equal togross tangible investments divided by the sum of gross tangible investments and the wagebill over the years 1995-2007. SKILL is equal to the number of non-production workersdivided by the number of production and non-production workers in 1998 (Istat data).R&D is the ratio of R&D expenditure to sales over the years 1997-1999.

Quality ladders. Quality ladders are taken from Khandelwal (2010). Starting from qualityladder measures at the ten-digit HS level, we first compute weighted averages of qualityladders at the six-digit HS level. The initial period total real value of the HS code is usedas a weight. Using a concordance table taken from Eurostat, we then compute weightedaverages of quality ladders at the four-digit CPA level. If a six-digit HS code was matchedwith two or more four-digit CPA codes, its value was assumed to be equally split acrossthe CPA codes. No suitable match was found for a group of HS codes amounting to about15 percent of total value in Khandelwal (2010)’s dataset (LAD).

Imports and exports. Italy’s imports and exports at current values by year, partner andproduct (Combined Nomenclature) over the years 1988-2006 are taken from the EurostatCOMEXT database. Data are then aggregated to the four-digit level of CPA classificationusing concordance tables available on the Eurostat website.

38

U.S. import penetration. U.S. trade data by year, partner and six-digit NAICS productare taken from Schott’s website. U.S. industry sales data at the six-digit NAICS levelare taken from the NBER productivity database. Data are then converted to the four-digit NACE rev.1 level, using a concordance table published by the Bureau of EconomicAnalysis and used to compute LWPENUS and OTPENUS .

Industrial production. Industrial production is taken from the IMF International Finan-cial Statistics (for advanced economies and China) and from the OECD Main EconomicIndicators (for other low-wage countries). Industrial production refers to the industrialsector for China and to the manufacturing sector for all the other countries. Low-wageindustrial production is interacted with an industry-level measure of routine and man-ual tasks intensity (IPLW ∗ ROUTINE). Industrial production in other countries isinteracted with an industry-level measure of non-routine and creative tasks intensity(IPOT ∗ NONROUTINE). Routine and manual tasks intensity is obtained as a com-posite of the following O*Net tasks: 4.C.3.d.3 Pace determined by speed of equipment;4.A.3.a.3 Controlling machines and processes; 4.C.2.d.1.i Spend time making repetitivemotions; 4.C.3.b.7 Importance of repeating the same tasks; 4.C.2.d.1.g Spend time usinghands to handle, control or feel objects, tools or controls. Non-routine and creative tasksintensity is obtained as a composite of the following O*Net tasks: 4.A.2.a.4 Analyzingdata/information; 4.A.2.b.2 Thinking creatively; 4.A.4.a.1 Interpreting information forothers. The O*Net database reports information on tasks by occupational profile. Foreach six-digit NAICS sector, a weighted average of each task is computed using the shareof each occupational profile in the sector’s total employment. Data are then aggregated atthe four-digit NAICS sector and converted to the four-digit NACE rev.1 level. The com-posite measure, obtained as the sum of the various task measures, is then standardized toa variable with zero mean and standard deviation equal to one.

Transport costs. Transport costs are computed as the ratio between freight costs and CIFimports for each US six-digit NAICS sector and country group (using data available fromSchott’s website). Data are then converted to the four-digit NACE rev.1 level, using aconcordance table published by the Bureau of Economic Analysis. We compute transportcosts separately for low-wage countries (TRANSPORTLW ) and for the other countries(TRANSPORTOT ).

Tariffs. Tariffs are computed as the ratio between duties and CIF imports for each U.S.six-digit NAICS sector and country group (using data available from Schott’s website).Data are then converted to the four-digit NACE rev.1 level, using a concordance tablepublished by the Bureau of Economic Analysis. We compute tariffs separately for low-wage countries (TARIFFLW ) and for the other countries (TARIFFOT ).

Exchange rates. Nominal exchange rates (annual averages) are taken from the Bank ofItaly. The exchange rate index is defined as a geometric average of foreign countries’nominal exchange rates. The weights are the share of each foreign country’s goods inItaly’s total imports in 1997. Weighted exchange rates are computed separately for low-wage countries (EXCHLW ) and for the rest of the world (EXCHOT ).

Relative export unit values. Export unit values are computed using the BACI tradedatabase available from Cepii (Gaulier and Zignago 2010). The database reports har-monized trade values (in dollars) and quantities (in tons), by reporting country, partner

39

country, HS six-digit product and year between 1995 and 2007. Data are converted tofour-digit NACE rev. 1 classification using a concordance table provided by Eurostat.Unit values are computed as values divided by quantities. Relative export unit values arethe log of Italy’s export unit values minus low-wage countries’ unit values, where low-wagecountries are defined as in the rest of the paper.

Input-output coefficients. Input-output coefficients for the upstream and downstream sec-tors are computed using the symmetric activity-by-activity table for Germany’s total pro-duction in the year 2000 (source: Eurostat).

Two-digit sector codes. 15 Food and beverages. 17 Textiles. 18 Wearing apparel. 19Leather products. 20 Wood and wood products. 21 Pulp, paper and printing. 22 Publish-ing and printing. 23 Coke and petroleum products. 24 Chemicals and chemical products.25 Rubber and plastic products. 26 Non-metallic mineral products. 27 Basic metals. 28Fabricated metal products. 29 Machinery and equipment. 31 Electrical machinery. 32Radio, television and communication machinery. 33 Medical, precision and optical instru-ments. 34 Motor vehicles. 35 Other transport equipment. 36 Furniture, toys and othermanufacturing products.

Macro sector aggregations. Textiles, Wearing apparel and Leather products (17, 18, 19).Food and beverages, Wood and wood products, Pulp, paper and printing, Publishing andprinting, Non-metallic mineral products, Furniture, toys and other manufacturing prod-ucts (15, 20, 21, 22, 26, 36). Coke and petroleum products, Chemicals and chemicalproducts, Rubber and plastic products (23, 24, 25). Basic metals, Fabricated metal prod-ucts, Machinery and equipment, Motor vehicles, Other transport equipment (27, 28, 29,34, 35). Electrical machinery, Radio, television and communication machinery, Medical,precision and optical instruments (31, 32, 33).

40

References

Acemoglu D., Autor D.H., 2010. Skills, Tasks and Technologies: Implicationsfor Employment and Earnings, in Ashenfelter O., Card D.E.. Handbookof Labor Economics, vol. 4, Elsevier, Amsterdam.

Altomonte C., Barattieri A., Rungi A., 2008. Import Penetration,Intermediate Inputs and Productivity: Evidence from Italian Firms.DYNREG Discussion Papers no. 23.

Amiti M., Davis D.R., 2008. Trade, Firms, and Wages: Theory and Evidence.National Bureau of Economic Research Working Paper no. 14106.

Amiti M., Konings J., 2007. Trade Liberalization, Intermediate Inputs, andProductivity: Evidence from Indonesia. American Economic Review97(5), 1611-1638.

Arkolakis C., Costinot A., Rodrıguez-Clare A., 2012. New Trade Models,Same Old Gains?. American Economic Review 102(1), 94-130.

Auer R., Fischer A.M., 2010. The Effect of Low-Wage Import Competitionon U.S. Inflationary Pressure, Journal of Monetary Economics 57(4),491-503.

Autor D.H., Dorn D., Hanson G., 2011. The China Syndrome: LocalLabor Market Effects of Import Competition in the United States, MITWorking Paper.

Baldwin J.R., Lileeva A., 2008. Adapting to Import Competition: Effects ofLow-wage Trade on Commodity Mix in Canadian Manufacturing Plants.Economic Analysis Research Paper, Statistics Canada no. 54.

Bernard A.B., Jensen J.B., Schott P.K., 2002. Survival of the Best Fit:Competition from Low-wage Countries and the (Uneven) Growth of U.S.Manufacturing Plants. National Bureau of Economic Research WorkingPaper no. 9170.

Bernard A.B., Jensen J.B., Schott P.K., 2006. Survival of the BestFit: Exposure to Low-wage Countries and the (Uneven) Growth ofU.S. Manufacturing Plants. Journal of International Economics 68(1),219-237.

Bloom N., Draca M., Van Reenen J., 2011. Trade Induced TechnicalChange: The Impact of Chinese Imports on Innovation, Diffusion andProductivity. Mimeo.

41

Bugamelli M., Fabiani S., Sette E., 2010. The Pro-Competitive Effect ofImports from China: an Analysis on Firm-Level Price Data, Temi didiscussione del Servizio Studi, Banca d’Italia no. 737.

Bugamelli M., Rosolia A., 2006. Produttivita e concorrenza estera. Temi didiscussione del Servizio Studi, Banca d’Italia no. 578.

Costinot A., Vogel J., 2010. Matching and Inequality in the World Economy.Journal of Political Economy 118(4), 747-786.

Eaton J., Kortum S., 2002. Technology, Geography, and Trade. Econometrica70(5), 1741-1779.

Feenstra R.C., 1994. New Product Varieties and the Measurement ofInternational Prices. American Economic Review 84(1), 157-177.

Feenstra R.C., 1995. Estimating the Effects of Trade Policy, in GrossmanG., Rogoff K. (eds.) Handbook of International Economics III. ElsevierScience B.V..

Fontagne L., Gaulier G., Zignago S., 2008. Specialization across Varieties andNorth-South Competition. Economic Policy 23, 51-91.

Freeman R., Katz L., 1991. Industrial Wage and Employment Determinationin an Open Economy, in Abowd J.M., Freeman R. (eds.) Immigration,Trade and Labor Market. University of Chicago Press.

Gaulier G., Zignago S., 2010. BACI: International Trade Database at theProduct-level: The 1994-2007 Version. CEPII Document de travail, no.23.

Grossman G., 1982. The Employment and Wage Effects of ImportCompetition in the United States. National Bureau Working Paper no.1041.

Istat, 2009. Conti economici delle imprese: Anno 2005. Rome.

Kandilov I., 2009. The Effects of Trade with Developing Countries on theRegional Demand for Skill in the U.S.: Evidence from County Data.Journal of Regional Science 49(3), 459482.

Kane T.J., Rouse C.E., Staiger D., 1999. Estimating Returns to SchoolingWhen Schooling is Misreported. National Bureau Working Paper no.7235.

Khandelwal A., 2010. The Long and Short (of) Quality Ladders. Review ofEconomic Studies 77, 1450-1476.

Krugman P., 2008. Trade and Wages, Reconsidered. Brookings Papers onEconomic Activity no. 1, 103-154.

42

Krugman P., Lawrence R., 1993. Trade, Jobs and Wages. National BureauWorking Paper no. 4478.

Lawrence C., Spiller P.T., 1983. Product Diversity, Economies of Scale, andInternational Trade. The Quarterly Journal of Economics 98(1), 63-83.

Long J.B., Plosser C. I., 1983. Real Business Cycles. The Journal of PoliticalEconomy 91(1), 39-69.

OECD, 2006. Structural and Demographic Business Statistics. Paris.

Pindyck R.S., Rotemberg J.J., 1987. Are Imports to Blame? Attribution ofInjury under the 1974 Trade Act. Journal of Law and Economics 30(1),101-122.

Revenga A., 1992. Exporting Jobs? The Impact of Import Competition onEmployment and Wages in U.S. Manufacturing. The Quarterly Journalof Economics 107(1), 255-284.

Sachs J.D., Shatz H.J., 1994. Trade and Jobs in U.S. Manufacturing.Brookings Papers on Economic Activity no. 1, 1-84.

Schott P.J., 2004. Across-Product versus Within-Product Specializationin International Trade. The Quarterly Journal of Economics 119(2),647-678.

Schott P.J., 2008. The Relative Sophistication of Chinese Exports. EconomicPolicy 53, 5-49.

Shea J., 2002. Complementarities and Comovements. Journal of Money,Credit and Banking 34(2), 412-433.

Smarzynska Javorcik, B., 2004. Does Foreign Direct Investment Increasethe Productivity of Domestic Firms? In Search of Spillovers ThroughBackward Linkages. American Economic Review 94, 605-27.

Topalova P., Khandelwal A., 2011. Trade Liberalization and FirmProductivity: The Case of India, The Review of Economics and Statistics93(3), 995-1009.

United Nations, 2009. International Trade Statistics Yearbook: 2008. NewYork.

White R., 2008. Import Source Reallocation and U.S. ManufacturingEmployment, 1972–2001. Open Economies Review 19(3), 403-410.

43

Table 1: SBS and national accounts data

Employment Value addedSBS NA % SBS NA %

1995 4771 5065 94.2 164 190 86.51996 4866 5005 97.2 186 197 94.21997 4807 4985 96.4 185 204 90.61998 4852 5071 95.7 191 205 93.21999 4774 5033 94.9 186 206 90.32000 4809 5005 96.1 204 216 94.32001 4820 4994 96.5 202 220 91.82002 4766 5034 94.7 202 222 91.02003 4749 5072 93.6 201 221 90.82004 4647 5026 92.5 206 227 90.62005 4569 4990 91.6 207 227 90.92006 4550 5034 90.4 217 237 91.82007 4576 5069 90.3 232 251 92.2Average 4735 5030 94.2 199 217 91.4

Data for Italy’s manufacturing sector. Employment is in thousands of persons, value added(at factor costs) in EUR billions at current prices.

Table 2: Summary statistics

Mean Std. Dev. Min. Max. NLWPEN 0.03 0.07 0 1.47 2454OTPEN 0.25 0.24 0 1.45 2454Employment (persons) 21133 28976 10 206838 2454Number of firms (units) 2408 5059 3 39676 2454Value added (EUR mln) 908 1091 0 9353 2454Sales (EUR mln) 3626 4976 0 65115 2447Wage bill (EUR mln) 379 466 0 3841 2454Skill intensity (%) 29.3 12.9 14.1 58.9 2454Capital intensity (%) 23.1 7.5 8.6 51.8 2454R&D intensity (%) 0.5 1.6 0 15.9 2435Ladder length (log) 2.0 0.8 -0.1 4.6 1899

44

Table 3: OLS estimates

(1) (2) (3) (4) (5) (6)

LWPENt−2 -1.766*** -0.580*** -0.434*** 0.028 -2.121*** -1.243***(0.251) (0.200) (0.163) (0.116) (0.403) (0.390)

OTPENt−2 -0.005 0.026 -0.028 -0.065 0.157 0.225(0.086) (0.079) (0.063) (0.069) (0.219) (0.222)

Estimation Fixed Fixed First First Long Longeffects effects diff. diff. diff. diff.

4d sector FE Yes Yes - - - -2d sector FE - - - - - Yes2d sector*year No Yes No Yes No NoYear Yes Yes Yes Yes No NoR2 0.143 0.366 0.033 0.210 0.226 0.503Observations 2454 2454 2214 2214 217 217Sectors 227 227 227 227 217 217

OLS regressions. The dependent variable is log employment in a four-digit sector.LWPENt−2 is low-wage import penetration, OTPENt−2 is import penetration fromthe rest of the world. Columns (1)-(2) are estimated with four-digit sector fixed effects,columns (3)-(4) are estimated with first differences, columns (5)-(6) are estimated withlong difference (changes between 1996 and 2007). Columns (1)-(4) include year fixed ef-fects. Columns (2) and (4) include the interactions between two-digit sector dummies andyear dummies. Column (6) includes two-digit sector dummies. Robust standard errors.Significance: * 0.10, ** 0.05, *** 0.01.

45

Table 4: IV estimates: first-stage (LWPEN)

Instrument US IP*tasks Trade costs(1) (2) (3) (4) (5) (6)

LWPENUS 0.393*** 0.377***(0.026) (0.035)

OTPENUS -0.018 -0.021**(0.012) (0.010)

IPLW ∗ROUTINE 0.008* -0.007(0.005) (0.009)

IPOT ∗NONROUTINE -0.124*** -0.026(0.019) (0.023)

TARIFFLW -0.083 0.274**(0.139) (0.111)

TARIFFOT -0.557*** -0.014(0.135) (0.088)

TRANSPORTLW -0.137*** -0.045*(0.038) (0.027)

TRANSPORTOT 0.064 -0.020(0.092) (0.080)

EXCHLW -0.001 -0.001(0.001) (0.001)

EXCHOT 0.092*** 0.022(0.017) (0.019)

F 113.2 58.2 26.2 2.6 9.8 1.94d sector FE Yes Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes Yes2d sector*year FE No Yes No Yes No YesObservations 2224 2224 2339 2339 2120 2120Sectors 205 205 216 216 196 196

First-stage results of IV regressions for the endogenous variable LWPENt−2. The instruments are: the corre-sponding import penetration measures for the U.S. (LWPENUS and OTPENUS) in columns (1)-(2); the inter-action between industrial production in low-wage countries and routine tasks intensity (IPLW ∗ ROUTINE)and the interaction between industrial production in advanced countries and non-routine tasks intensity(IPOT ∗ NONROUTINE) in columns (3)-(4); tariffs, transport costs and exchange rates (TARIFFLW ,TARIFFOT , TRANSPORTLW , TRANSPORTOT , EXCHLW , EXCHOT ) in columns (5)-(6). All columnsinclude year and four-digit sector fixed effects. Columns (2), (4) and (6) include the interactions between two-digit sector dummies and year dummies. Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

46

Table 5: IV estimates: first-stage (OTPEN)

Instrument US IP*tasks Trade costs(1) (2) (3) (4) (5) (6)

LWPENUS -0.051 0.105(0.061) (0.072)

OTPENUS 0.380*** 0.299***(0.056) (0.036)

IPLW ∗ROUTINE -0.023 -0.033*(0.015) (0.019)

IPOT ∗NONROUTINE 0.061 0.147***(0.054) (0.052)

TARIFFLW 0.128 0.095(0.367) (0.333)

TARIFFOT -0.048 0.450**(0.301) (0.190)

TRANSPORTLW -0.335** -0.232*(0.141) (0.119)

TRANSPORTOT -0.922*** -0.185(0.289) (0.196)

EXCHLW -0.010** 0.001(0.004) (0.004)

EXCHOT -0.122** 0.016(0.059) (0.038)

F 23.1 36.9 3.5 1.4 5.6 2.74d sector FE Yes Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes Yes2d sector*year FE No Yes No Yes No YesObservations 2224 2224 2339 2339 2120 2120Sectors 205 205 216 216 196 196

First-stage results of IV regressions for the endogenous variable OTWPENt−2. The instruments are: the corre-sponding import penetration measures for the U.S. (LWPENUS and OTPENUS) in columns (1)-(2); the inter-action between industrial production in low-wage countries and routine tasks intensity (IPLW ∗ ROUTINE)and the interaction between industrial production in advanced countries and non-routine tasks intensity(IPOT ∗ NONROUTINE) in columns (3)-(4); tariffs, transport costs and exchange rates (TARIFFLW ,TARIFFOT , TRANSPORTLW , TRANSPORTOT , EXCHLW , EXCHOT ) in columns (5)-(6). All columnsinclude year and four-digit sector fixed effects. Columns (2), (4) and (6) include the interactions between two-digit sector dummies and year dummies. Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

47

Table 6: IV estimates: second-stage

Instrument US IP*tasks Trade costs(1) (2) (3) (4) (5) (6)

LWPENt−2 -2.914*** -1.886*** -4.218*** -8.004* -4.878*** -3.538(0.283) (0.306) (1.443) (4.745) (0.635) (2.805)

OTPENt−2 -0.030 -0.334* -2.981 -1.599 -0.079 1.612*(0.149) (0.198) (2.303) (1.738) (0.318) (0.929)

Kleibergen-Paap F 22.7 36.4 0.8 0.9 5.1 1.5Hansen J - - - - 4.7 12.34d sector FE Yes Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes Yes2d sector*year FE No Yes No Yes No YesObservations 2224 2224 2339 2339 2120 2120Sectors 205 205 216 216 196 196

Second-stage results of IV regressions. The dependent variable is log employment in a four-digit sec-tor. LWPENt−2 is low-wage import penetration, OTPENt−2 is import penetration from the rest of theworld. The instruments are: the corresponding import penetration measures for the U.S. (LWPENUS

and OTPENUS) in columns (1)-(2); the interaction between industrial production in low-wage countriesand routine tasks intensity (IPLW ∗ ROUTINE) and the interaction between industrial production inadvanced countries and non-routine tasks intensity (IPOT ∗NONROUTINE) in columns (3)-(4); tariffs,transport costs and exchange rates (TARIFFLW , TARIFFOT , TRANSPORTLW , TRANSPORTOT ,EXCHLW , EXCHOT ) in columns (5)-(6). All columns include year and four-digit sector fixed effects.Columns (2), (4) and (6) include the interactions between two-digit sector dummies and year dummies.Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

48

Table 7: Alternative dependent variables

(1) (2) (3) (4) (5)Empl. Value add. Sales Wage No. of

bill firms

LWPENt−2 -0.580*** -0.860*** -0.809*** -0.866*** -0.502***(0.200) (0.266) (0.241) (0.250) (0.182)

OTPENt−2 0.026 0.178 0.118 0.028 0.176***(0.079) (0.119) (0.108) (0.097) (0.067)

R2 0.366 0.291 0.299 0.335 0.3364d sector FE Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes2d sector*year FE Yes Yes Yes Yes YesObservations 2454 2401 2394 2399 2454Sectors 227 227 227 227 227

OLS regressions. The dependent variable is log employment in a four-digit sector in col-umn (1), log deflated value added in column (2), log deflated sales in column (3), logdeflated wage bill in column (4) and log number of firms in column (5). LWPENt−2

is low-wage import penetration, OTPENt−2 is import penetration from the rest of theworld. All columns include year and four-digit sector fixed effects and the interactionsbetween two-digit sector dummies and year dummies. Robust standard errors. Signifi-cance: * 0.10, ** 0.05, *** 0.01.

49

Table 8: Alternative lags and low-wage country thresholds

Lags Low-wage country thresholds1 year 3 year 4 year 5% 20% 30%

LWPENt−lag -0.709*** -0.580** -0.805***(0.173) (0.240) (0.190)

OTPENt−lag 0.091 -0.037 -0.105(0.075) (0.074) (0.067)

LWPENthreshold -0.465** -0.385*** -0.432***(0.226) (0.133) (0.114)

OTPENthreshold -0.023 0.037 0.079(0.073) (0.083) (0.088)

4d sector FE Yes Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes Yes2d sector*year FE Yes Yes Yes Yes Yes YesR2 0.370 0.363 0.359 0.363 0.365 0.368Observations 2676 2231 2006 2454 2454 2454Sectors 227 227 227 227 227 227

OLS regressions. The dependent variable is log employment in a four-digit sector. LWPEN is low-wage importpenetration, OTPEN is import penetration from the rest of the world. They are included with a 1-year lag incolumn (1), 3-year lag in column (2) and 4-year lag in column (3). They are computed using a 5% threshold(relative to U.S. GDP per capita at current prices in 2006) in column (4), 20% threshold in column (5) and 30%threshold in column (6). All columns include year and four-digit sector fixed effects and the interactions betweentwo-digit sector dummies and year dummies. Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

50

Table 9: Factor intensity and quality ladder

(1) (2) (3) (4) (5)

LWPENt−2 0.038 -0.493*** -0.627*** -0.423** 0.114(0.207) (0.188) (0.207) (0.185) (0.231)

LWPENt−2 ∗ SKILL 0.090*** 0.080***(0.017) (0.023)

LWPENt−2 ∗ CAP 0.017 -0.000(0.018) (0.023)

LWPENt−2 ∗R&D 0.306*** 0.016(0.101) (0.132)

LWPENt−2 ∗ LAD 0.586*** 0.458***(0.096) (0.075)

OTPENt−2 0.043 0.026 0.055 0.051 0.063(0.074) (0.079) (0.074) (0.082) (0.078)

4d sector FE Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes2d sector*year FE Yes Yes Yes Yes YesR2 0.378 0.366 0.376 0.366 0.384Observations 2454 2454 2435 1899 1890Sectors 227 227 224 174 173Effect of LWPEN on employment at selected values of interacted variable

SKILL CAP R&D LAD -Mean - sd -1.105*** -0.622*** -1.119*** -0.875*** -Mean + sd 1.181*** -0.363 -0.136 0.029 -

OLS regressions. The dependent variable is log employment in a four-digit sector. LWPENt−2 is low-wage import penetration, OTPENt−2 is import penetration from the rest of the world. SKILL is non-production workers ratio in a four-digit sector, CAP is capital intensity, R&D is R&D expenditure onsales and LAD is quality ladder length. SKILL, CAP , R&D and LAD are centered with mean zero.All columns include year and four-digit sector fixed effects and the interactions between two-digit sectordummies and year dummies. The lower panel reports the estimated coefficient of LWPENt−2 on employ-ment at selected values of the interacted variable (mean minus one standard deviation and mean plus onestandard deviation. Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

51

Table 10: Quality ladder and export unit values

(1) (2) (3) (4)

LWPENt−2 0.567 -1.046 0.631 -1.088(0.517) (0.924) (0.521) (0.953)

LWPENt−2 ∗ LAD 0.867** 0.916**(0.374) (0.383)

OTPENt−2 -0.001 -0.017 -0.003 0.089(0.082) (0.242) (0.083) (0.252)

OTPENt−2 ∗ LAD 0.005 -0.044(0.112) (0.114)

4d sector FE Yes Yes No No6d product FE No No Yes Yes2d sector*year Yes Yes Yes YesYear Yes Yes Yes YesR2 0.170 0.170 0.551 0.555Observations 44312 44312 44312 44312

OLS regressions. The dependent variable is log Italy’s export unit values relative tolow-wage countries’ export unit values in a six-digit product. LWPENt−2 is low-wageimport penetration in the corresponding four-digit sector, LWPENt−2 ∗LAD is its inter-action with quality ladder, OTPENt−2 is import penetration from the rest of the world,OTPENt−2 ∗ LAD is its interaction with quality ladder. Columns (1)-(2) are estimatedwith four-digit sector fixed effects, columns (3)-(4) are estimated with six-digit productfixed effects. All columns include year dummies and the interactions between two-digitsector dummies and year dummies. Standard errors are clustered at the four-digit level.Significance: * 0.10, ** 0.05, *** 0.01.

52

Table 11: Two-digit sectors: OLS

(1) (2) (3) (4) (5) (6)

LWPENt−2 -3.677*** -1.043*** -3.951*** -1.596*** -3.525*** -2.436***(0.303) (0.347) (0.389) (0.467) (0.329) (0.555)

OTPENt−2 -0.225** -0.334*** -0.357*** -0.322*** -0.242*** -0.140(0.095) (0.094) (0.082) (0.099) (0.078) (0.106)

DOWNLWPENt−2 -5.103*** -0.897 -6.587*** -4.491***(0.923) (1.183) (0.892) (1.410)

DOWNOTPENt−2 0.901*** 0.834***(0.202) (0.294)

UPLWPENt−2 4.367*** 3.576** 2.864** 1.816(1.321) (1.546) (1.257) (1.479)

UPOTPENt−2 0.064 0.366(0.377) (0.349)

R2 0.524 0.734 0.641 0.751 0.689 0.7692d sector FE Yes Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes YesMacro-sector*year FE No Yes No Yes No YesObservations 231 231 231 231 231 231Sectors 21 21 21 21 21 21

OLS regressions. The dependent variable is log employment in a two-digit sector. LWPENt−2, UPLWPENt−2

and DOWNLWPENt−2 are low-wage import penetration in the same sector, in upstream sectors and in down-stream sectors, respectively. OTPENt−2, UPOTPENt−2 and DOWNOTPENt−2 are import penetration fromthe rest of the world in the same sector, in upstream sectors and in downstream sectors, respectively. Importpenetration in upstream and downstream sectors is a weighted average across all sectors (weights are given byinput-output coefficients). All columns include year and two-digit sector fixed effects. Columns (4)-(6) includethe interactions between macro sector dummies and year dummies. Macro sectors are defined as aggregations oftwo-digit sectors (see Appendix). Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

53

Table 12: Two-digit sectors: IV

(1) (2) (3) (4) (5) (6)

LWPENt−2 -3.926*** -0.759 -4.538*** -2.295** -4.252*** -4.870***(0.311) (0.517) (0.427) (0.918) (0.445) (1.524)

OTPENt−2 0.024 -0.341* -0.207 -0.318 -0.141 0.076(0.189) (0.200) (0.162) (0.206) (0.162) (0.336)

DOWNLWPENt−2 -4.798*** -1.582 -5.502*** -6.133**(0.947) (1.410) (1.378) (2.464)

DOWNOTPENt−2 0.985*** 1.549**(0.274) (0.709)

UPLWPENt−2 5.339*** 4.646 4.736** 4.785(1.877) (3.085) (2.227) (3.400)

UPOTPENt−2 -0.217 -0.253(0.514) (0.624)

Cragg-Donald F statistic 52.7 31.0 25.2 15.1 15.1 3.5R2 0.502 0.733 0.625 0.746 0.671 0.7292d sector FE Yes Yes Yes Yes Yes YesYear FE Yes Yes Yes Yes Yes YesMacro-sector*year FE No Yes No Yes No YesObservations 231 231 231 231 231 231Sectors 21 21 21 21 21 21

Second-stage results of IV regressions. The dependent variable is log employment in a two-digit sector.LWPENt−2, UPLWPENt−2 and DOWNLWPENt−2 are low-wage import penetration in the same sector, inupstream sectors and in downstream sectors, respectively. OTPENt−2, UPOTPENt−2 and DOWNOTPENt−2

are import penetration from the rest of the world in the same sector, in upstream sectors and in downstreamsectors, respectively. Import penetration in upstream and downstream sectors is a weighted average across allsectors (weights are given by input-output coefficients). All columns include year and two-digit sector fixed effects.Columns (4)-(6) include the interactions between macro sector dummies and year dummies. Macro sectors aredefined as aggregations of two-digit sectors (see Appendix). The instruments are the corresponding import pene-tration measures for the U.S. (LWPENUS , UPLWPENUS , DOWNLWPENUS , OTPENUS , UPOTPENUS

and DOWNOTPENUS). Robust standard errors. Significance: * 0.10, ** 0.05, *** 0.01.

54

Figure 1: Change in log employment and change in low-wage importpenetration (1995-2007)

55

Figure 2: Employment and low-wage import penetration in up-stream and downstream sectors (1995-2007)

56

(*) Requestsforcopiesshouldbesentto:Bancad’Italia–ServizioStudidistrutturaeconomicaefinanziaria–DivisioneBibliotecaeArchiviostorico–ViaNazionale,91–00184Rome–(fax00390647922059).TheyareavailableontheInternetwww.bancaditalia.it.

RECENTLYPUBLISHED“TEMI”(*)

N. 856 – Optimal dynamic public communication,byMarcelloMiccoli(February2012).

N. 857 – Modelling loans to non-financial corporations in the euro area,byChristofferKokSørensen,DavidMarquésIbáñezandCarlottaRossi(February2012).

N. 858 – Economic (in)stability under monetary targeting,byLucaSessa(March2012).

N. 859 – Optimal fiscal policy when agents fear government default,byFrancescoCaprioli,PietroRizzaandPietroTommasino(March2012).

N. 860 – The effects of unemployment benefits in Italy: evidence from an institutional

change,byAlfonsoRosoliaandPaoloSestito(March2012).

N. 861 – Monetary policy and the flow of funds in the euro area,byRiccardoBonci(March2012).

N. 862 – Does gender matter for public spending? Empirical evidence from Italian

municipalities,byMassimilianoRigonandGiuliaM.Tanzi(April2012).

N. 863 – House price cycles in emerging economies,byAlessioCiarlone(April2012).

N. 864 – Measuring the underground economy with the currency demand approach: a

reinterpretation of the methodology, with an application to Italy,byGuerinoArdizzi,CarmeloPetraglia,MassimilianoPiacenzaandGilbertoTurati(April2012).

N. 865 – Corporate geography in multi-plant firms,byRitaCappariello,StefanoFedericoandRobertaZizza(April2012).

N. 866 – Don’t stand so close to me: the urban impact of immigration,byAntonioAccetturo,FrancescoManaresi,SauroMocettiandElisabettaOlivieri(April2012).

N. 867 – Disinflation effects in a medium-scale New Keynesian model: money supply rule

versus interest rate rule,byGuidoAscariandTizianoRopele(April2012).

N. 868 – The economic costs of organized crime: evidence from southern Italy, byPaoloPinotti(April2012).

N. 869 – Network effects of public transport infrastructure: evidence on Italian regions,byValterDiGiacinto,GiacintoMicucciandPasqualinoMontanaro(July2012).

N. 870 – To misreport or not to report? The measurement of household financial wealth,byAndreaNeriandMariaGiovannaRanalli(July2012).

N. 871 – Capital destruction, jobless recoveries, and the discipline device role of

unemployment,byMariannaRiggi(July2012).

N. 872 – Selecting predictors by using Bayesian model averaging in bridge models, byLorenzoBencivelli,MassimilianoMarcellinoandGianlucaMoretti(July2012).

N. 873 – Euro area and global oil shocks: an empirical model-based analysis,byLorenzoForni,AndreaGerali,AlessandroNotarpietroandMassimilianoPisani(July2012).

N. 874 – Evidence on the impact of R&D and ICT investment on innovation and productivity

in Italian firms,byBronwynH.Hall,FrancescaLottiandJacquesMairesse(July2012).

N. 875 – Family background, self-confidence and economic outcomes,byAntonioFilippinandMarcoPaccagnella(July2012).

N. 876 – Banks’ reactions to Basel-III,byPaoloAngeliniandAndreaGerali(July2012).

"TEMI" LATER PUBLISHED ELSEWHERE

2009

F. PANETTA, F. SCHIVARDI and M. SHUM, Do mergers improve information? Evidence from the loan market, Journal of Money, Credit, and Banking, v. 41, 4, pp. 673-709, TD No. 521 (October 2004).

M. BUGAMELLI and F. PATERNÒ, Do workers’ remittances reduce the probability of current account reversals?, World Development, v. 37, 12, pp. 1821-1838, TD No. 573 (January 2006).

P. PAGANO and M. PISANI, Risk-adjusted forecasts of oil prices, The B.E. Journal of Macroeconomics, v. 9, 1, Article 24, TD No. 585 (March 2006).

M. PERICOLI and M. SBRACIA, The CAPM and the risk appetite index: theoretical differences, empirical similarities, and implementation problems, International Finance, v. 12, 2, pp. 123-150, TD No. 586 (March 2006).

R. BRONZINI and P. PISELLI, Determinants of long-run regional productivity with geographical spillovers: the role of R&D, human capital and public infrastructure, Regional Science and Urban Economics, v. 39, 2, pp.187-199, TD No. 597 (September 2006).

U. ALBERTAZZI and L. GAMBACORTA, Bank profitability and the business cycle, Journal of Financial Stability, v. 5, 4, pp. 393-409, TD No. 601 (September 2006).

F. BALASSONE, D. FRANCO and S. ZOTTERI, The reliability of EMU fiscal indicators: risks and safeguards, in M. Larch and J. Nogueira Martins (eds.), Fiscal Policy Making in the European Union: an Assessment of Current Practice and Challenges, London, Routledge, TD No. 633 (June 2007).

A. CIARLONE, P. PISELLI and G. TREBESCHI, Emerging Markets' Spreads and Global Financial Conditions, Journal of International Financial Markets, Institutions & Money, v. 19, 2, pp. 222-239, TD No. 637 (June 2007).

S. MAGRI, The financing of small innovative firms: the Italian case, Economics of Innovation and New Technology, v. 18, 2, pp. 181-204, TD No. 640 (September 2007).

V. DI GIACINTO and G. MICUCCI, The producer service sector in Italy: long-term growth and its local determinants, Spatial Economic Analysis, Vol. 4, No. 4, pp. 391-425, TD No. 643 (September 2007).

F. LORENZO, L. MONTEFORTE and L. SESSA, The general equilibrium effects of fiscal policy: estimates for the euro area, Journal of Public Economics, v. 93, 3-4, pp. 559-585, TD No. 652 (November 2007).

Y. ALTUNBAS, L. GAMBACORTA and D. MARQUÉS, Securitisation and the bank lending channel, European Economic Review, v. 53, 8, pp. 996-1009, TD No. 653 (November 2007).

R. GOLINELLI and S. MOMIGLIANO, The Cyclical Reaction of Fiscal Policies in the Euro Area. A Critical Survey of Empirical Research, Fiscal Studies, v. 30, 1, pp. 39-72, TD No. 654 (January 2008).

P. DEL GIOVANE, S. FABIANI and R. SABBATINI, What’s behind “Inflation Perceptions”? A survey-based analysis of Italian consumers, Giornale degli Economisti e Annali di Economia, v. 68, 1, pp. 25-52, TD No. 655 (January 2008).

F. MACCHERONI, M. MARINACCI, A. RUSTICHINI and M. TABOGA, Portfolio selection with monotone mean-variance preferences, Mathematical Finance, v. 19, 3, pp. 487-521, TD No. 664 (April 2008).

M. AFFINITO and M. PIAZZA, What are borders made of? An analysis of barriers to European banking integration, in P. Alessandrini, M. Fratianni and A. Zazzaro (eds.): The Changing Geography of Banking and Finance, Dordrecht Heidelberg London New York, Springer, TD No. 666 (April 2008).

A. BRANDOLINI, On applying synthetic indices of multidimensional well-being: health and income inequalities in France, Germany, Italy, and the United Kingdom, in R. Gotoh and P. Dumouchel (eds.), Against Injustice. The New Economics of Amartya Sen, Cambridge, Cambridge University Press, TD No. 668 (April 2008).

G. FERRERO and A. NOBILI, Futures contract rates as monetary policy forecasts, International Journal of Central Banking, v. 5, 2, pp. 109-145, TD No. 681 (June 2008).

P. CASADIO, M. LO CONTE and A. NERI, Balancing work and family in Italy: the new mothers’ employment decisions around childbearing, in T. Addabbo and G. Solinas (eds.), Non-Standard Employment and Qualità of Work, Physica-Verlag. A Sprinter Company, TD No. 684 (August 2008).

L. ARCIERO, C. BIANCOTTI, L. D'AURIZIO and C. IMPENNA, Exploring agent-based methods for the analysis of payment systems: A crisis model for StarLogo TNG, Journal of Artificial Societies and Social Simulation, v. 12, 1, TD No. 686 (August 2008).

A. CALZA and A. ZAGHINI, Nonlinearities in the dynamics of the euro area demand for M1, Macroeconomic Dynamics, v. 13, 1, pp. 1-19, TD No. 690 (September 2008).

L. FRANCESCO and A. SECCHI, Technological change and the households’ demand for currency, Journal of Monetary Economics, v. 56, 2, pp. 222-230, TD No. 697 (December 2008).

G. ASCARI and T. ROPELE, Trend inflation, taylor principle, and indeterminacy, Journal of Money, Credit and Banking, v. 41, 8, pp. 1557-1584, TD No. 708 (May 2007).

S. COLAROSSI and A. ZAGHINI, Gradualism, transparency and the improved operational framework: a look at overnight volatility transmission, International Finance, v. 12, 2, pp. 151-170, TD No. 710 (May 2009).

M. BUGAMELLI, F. SCHIVARDI and R. ZIZZA, The euro and firm restructuring, in A. Alesina e F. Giavazzi (eds): Europe and the Euro, Chicago, University of Chicago Press, TD No. 716 (June 2009).

B. HALL, F. LOTTI and J. MAIRESSE, Innovation and productivity in SMEs: empirical evidence for Italy, Small Business Economics, v. 33, 1, pp. 13-33, TD No. 718 (June 2009).

2010

A. PRATI and M. SBRACIA, Uncertainty and currency crises: evidence from survey data, Journal of Monetary Economics, v, 57, 6, pp. 668-681, TD No. 446 (July 2002).

L. MONTEFORTE and S. SIVIERO, The Economic Consequences of Euro Area Modelling Shortcuts, Applied Economics, v. 42, 19-21, pp. 2399-2415, TD No. 458 (December 2002).

S. MAGRI, Debt maturity choice of nonpublic Italian firms , Journal of Money, Credit, and Banking, v.42, 2-3, pp. 443-463, TD No. 574 (January 2006).

G. DE BLASIO and G. NUZZO, Historical traditions of civicness and local economic development, Journal of Regional Science, v. 50, 4, pp. 833-857, TD No. 591 (May 2006).

E. IOSSA and G. PALUMBO, Over-optimism and lender liability in the consumer credit market, Oxford Economic Papers, v. 62, 2, pp. 374-394, TD No. 598 (September 2006).

S. NERI and A. NOBILI, The transmission of US monetary policy to the euro area, International Finance, v. 13, 1, pp. 55-78, TD No. 606 (December 2006).

F. ALTISSIMO, R. CRISTADORO, M. FORNI, M. LIPPI and G. VERONESE, New Eurocoin: Tracking Economic Growth in Real Time, Review of Economics and Statistics, v. 92, 4, pp. 1024-1034, TD No. 631 (June 2007).

U. ALBERTAZZI and L. GAMBACORTA, Bank profitability and taxation, Journal of Banking and Finance, v. 34, 11, pp. 2801-2810, TD No. 649 (November 2007).

M. IACOVIELLO and S. NERI, Housing market spillovers: evidence from an estimated DSGE model, American Economic Journal: Macroeconomics, v. 2, 2, pp. 125-164, TD No. 659 (January 2008).

F. BALASSONE, F. MAURA and S. ZOTTERI, Cyclical asymmetry in fiscal variables in the EU, Empirica, TD No. 671, v. 37, 4, pp. 381-402 (June 2008).

F. D'AMURI, O. GIANMARCO I.P. and P. GIOVANNI, The labor market impact of immigration on the western german labor market in the 1990s, European Economic Review, v. 54, 4, pp. 550-570, TD No. 687 (August 2008).

A. ACCETTURO, Agglomeration and growth: the effects of commuting costs, Papers in Regional Science, v. 89, 1, pp. 173-190, TD No. 688 (September 2008).

S. NOBILI and G. PALAZZO, Explaining and forecasting bond risk premiums, Financial Analysts Journal, v. 66, 4, pp. 67-82, TD No. 689 (September 2008).

A. B. ATKINSON and A. BRANDOLINI, On analysing the world distribution of income, World Bank Economic Review , v. 24, 1 , pp. 1-37, TD No. 701 (January 2009).

R. CAPPARIELLO and R. ZIZZA, Dropping the Books and Working Off the Books, Labour, v. 24, 2, pp. 139-162 ,TD No. 702 (January 2009).

C. NICOLETTI and C. RONDINELLI, The (mis)specification of discrete duration models with unobserved heterogeneity: a Monte Carlo study, Journal of Econometrics, v. 159, 1, pp. 1-13, TD No. 705 (March 2009).

L. FORNI, A. GERALI and M. PISANI, Macroeconomic effects of greater competition in the service sector: the case of Italy, Macroeconomic Dynamics, v. 14, 5, pp. 677-708, TD No. 706 (March 2009).

V. DI GIACINTO, G. MICUCCI and P. MONTANARO, Dynamic macroeconomic effects of public capital: evidence from regional Italian data, Giornale degli economisti e annali di economia, v. 69, 1, pp. 29-66, TD No. 733 (November 2009).

F. COLUMBA, L. GAMBACORTA and P. E. MISTRULLI, Mutual Guarantee institutions and small business finance, Journal of Financial Stability, v. 6, 1, pp. 45-54, TD No. 735 (November 2009).

A. GERALI, S. NERI, L. SESSA and F. M. SIGNORETTI, Credit and banking in a DSGE model of the Euro Area, Journal of Money, Credit and Banking, v. 42, 6, pp. 107-141, TD No. 740 (January 2010).

M. AFFINITO and E. TAGLIAFERRI, Why do (or did?) banks securitize their loans? Evidence from Italy, Journal

of Financial Stability, v. 6, 4, pp. 189-202, TD No. 741 (January 2010).

S. FEDERICO, Outsourcing versus integration at home or abroad and firm heterogeneity, Empirica, v. 37, 1, pp. 47-63, TD No. 742 (February 2010).

V. DI GIACINTO, On vector autoregressive modeling in space and time, Journal of Geographical Systems, v. 12, 2, pp. 125-154, TD No. 746 (February 2010).

L. FORNI, A. GERALI and M. PISANI, The macroeconomics of fiscal consolidations in euro area countries, Journal of Economic Dynamics and Control, v. 34, 9, pp. 1791-1812, TD No. 747 (March 2010).

S. MOCETTI and C. PORELLO, How does immigration affect native internal mobility? new evidence from Italy, Regional Science and Urban Economics, v. 40, 6, pp. 427-439, TD No. 748 (March 2010).

A. DI CESARE and G. GUAZZAROTTI, An analysis of the determinants of credit default swap spread changes before and during the subprime financial turmoil, Journal of Current Issues in Finance, Business and Economics, v. 3, 4, pp., TD No. 749 (March 2010).

P. CIPOLLONE, P. MONTANARO and P. SESTITO, Value-added measures in Italian high schools: problems and findings, Giornale degli economisti e annali di economia, v. 69, 2, pp. 81-114, TD No. 754 (March 2010).

A. BRANDOLINI, S. MAGRI and T. M SMEEDING, Asset-based measurement of poverty, Journal of Policy Analysis and Management, v. 29, 2 , pp. 267-284, TD No. 755 (March 2010).

G. CAPPELLETTI, A Note on rationalizability and restrictions on beliefs, The B.E. Journal of Theoretical Economics, v. 10, 1, pp. 1-11,TD No. 757 (April 2010).

S. DI ADDARIO and D. VURI, Entrepreneurship and market size. the case of young college graduates in Italy, Labour Economics, v. 17, 5, pp. 848-858, TD No. 775 (September 2010).

A. CALZA and A. ZAGHINI, Sectoral money demand and the great disinflation in the US, Journal of Money, Credit, and Banking, v. 42, 8, pp. 1663-1678, TD No. 785 (January 2011).

2011

S. DI ADDARIO, Job search in thick markets, Journal of Urban Economics, v. 69, 3, pp. 303-318, TD No. 605 (December 2006).

F. SCHIVARDI and E. VIVIANO, Entry barriers in retail trade, Economic Journal, v. 121, 551, pp. 145-170, TD No. 616 (February 2007).

G. FERRERO, A. NOBILI and P. PASSIGLIA, Assessing excess liquidity in the Euro Area: the role of sectoral distribution of money, Applied Economics, v. 43, 23, pp. 3213-3230, TD No. 627 (April 2007).

P. E. MISTRULLI, Assessing financial contagion in the interbank market: maximun entropy versus observed interbank lending patterns, Journal of Banking & Finance, v. 35, 5, pp. 1114-1127, TD No. 641 (September 2007).

E. CIAPANNA, Directed matching with endogenous markov probability: clients or competitors?, The RAND Journal of Economics, v. 42, 1, pp. 92-120, TD No. 665 (April 2008).

M. BUGAMELLI and F. PATERNÒ, Output growth volatility and remittances, Economica, v. 78, 311, pp. 480-500, TD No. 673 (June 2008).

V. DI GIACINTO e M. PAGNINI, Local and global agglomeration patterns: two econometrics-based indicators, Regional Science and Urban Economics, v. 41, 3, pp. 266-280, TD No. 674 (June 2008).

G. BARONE and F. CINGANO, Service regulation and growth: evidence from OECD countries, Economic Journal, v. 121, 555, pp. 931-957, TD No. 675 (June 2008).

R. GIORDANO and P. TOMMASINO, What determines debt intolerance? The role of political and monetary institutions, European Journal of Political Economy, v. 27, 3, pp. 471-484, TD No. 700 (January 2009).

P. ANGELINI, A. NOBILI e C. PICILLO, The interbank market after August 2007: What has changed, and why?, Journal of Money, Credit and Banking, v. 43, 5, pp. 923-958, TD No. 731 (October 2009).

L. FORNI, A. GERALI and M. PISANI, The Macroeconomics of Fiscal Consolidation in a Monetary Union: the Case of Italy, in Luigi Paganetto (ed.), Recovery after the crisis. Perspectives and policies, VDM Verlag Dr. Muller, TD No. 747 (March 2010).

A. DI CESARE and G. GUAZZAROTTI, An analysis of the determinants of credit default swap changes before and during the subprime financial turmoil, in Barbara L. Campos and Janet P. Wilkins (eds.), The Financial Crisis: Issues in Business, Finance and Global Economics, New York, Nova Science Publishers, Inc., TD No. 749 (March 2010).

A. LEVY and A. ZAGHINI, The pricing of government guaranteed bank bonds, Banks and Bank Systems, v. 6, 3, pp. 16-24, TD No. 753 (March 2010).

G. GRANDE and I. VISCO, A public guarantee of a minimum return to defined contribution pension scheme members, The Journal of Risk, v. 13, 3, pp. 3-43, TD No. 762 (June 2010).

P. DEL GIOVANE, G. ERAMO and A. NOBILI, Disentangling demand and supply in credit developments: a survey-based analysis for Italy, Journal of Banking and Finance, v. 35, 10, pp. 2719-2732, TD No. 764 (June 2010).

G. BARONE and S. MOCETTI, With a little help from abroad: the effect of low-skilled immigration on the female labour supply, Labour Economics, v. 18, 5, pp. 664-675, TD No. 766 (July 2010).

A. FELETTIGH and S. FEDERICO, Measuring the price elasticity of import demand in the destination markets of italian exports, Economia e Politica Industriale, v. 38, 1, pp. 127-162, TD No. 776 (October 2010).

S. MAGRI and R. PICO, The rise of risk-based pricing of mortgage interest rates in Italy, Journal of Banking and Finance, v. 35, 5, pp. 1277-1290, TD No. 778 (October 2010).

M. TABOGA, Under/over-valuation of the stock market and cyclically adjusted earnings, International Finance, v. 14, 1, pp. 135-164, TD No. 780 (December 2010).

S. NERI, Housing, consumption and monetary policy: how different are the U.S. and the Euro area?, Journal of Banking and Finance, v.35, 11, pp. 3019-3041, TD No. 807 (April 2011).

V. CUCINIELLO, The welfare effect of foreign monetary conservatism with non-atomistic wage setters, Journal of Money, Credit and Banking, v. 43, 8, pp. 1719-1734, TD No. 810 (June 2011).

A. CALZA and A. ZAGHINI, welfare costs of inflation and the circulation of US currency abroad, The B.E. Journal of Macroeconomics, v. 11, 1, Art. 12, TD No. 812 (June 2011).

I. FAIELLA, La spesa energetica delle famiglie italiane, Energia, v. 32, 4, pp. 40-46, TD No. 822 (September 2011).

R. DE BONIS and A. SILVESTRINI, The effects of financial and real wealth on consumption: new evidence from OECD countries, Applied Financial Economics, v. 21, 5, pp. 409–425, TD No. 837 (November 2011).

2012

A. ACCETTURO and G. DE BLASIO, Policies for local development: an evaluation of Italy’s “Patti Territoriali”, Regional Science and Urban Economics, v. 42, 1-2, pp. 15-26, TD No. 789 (January 2006).

FORTHCOMING

M. BUGAMELLI and A. ROSOLIA, Produttività e concorrenza estera, Rivista di politica economica, TD No. 578 (February 2006).

F. CINGANO and A. ROSOLIA, People I know: job search and social networks, Journal of Labor Economics, TD No. 600 (September 2006).

S. MOCETTI, Educational choices and the selection process before and after compulsory school, Education Economics, TD No. 691 (September 2008).

P. SESTITO and E. VIVIANO, Reservation wages: explaining some puzzling regional patterns, Labour, TD No. 696 (December 2008).

P. PINOTTI, M. BIANCHI and P. BUONANNO, Do immigrants cause crime?, Journal of the European Economic Association, TD No. 698 (December 2008).

F. LIPPI and A. NOBILI, Oil and the macroeconomy: a quantitative structural analysis, Journal of European Economic Association, TD No. 704 (March 2009).

F. CINGANO and P. PINOTTI, Politicians at work. The private returns and social costs of political connections, Journal of the European Economic Association, TD No. 709 (May 2009).

Y. ALTUNBAS, L. GAMBACORTA, and D. MARQUÉS-IBÁÑEZ, Bank risk and monetary policy, Journal of Financial Stability, TD No. 712 (May 2009).

G. BARONE and S. MOCETTI, Tax morale and public spending inefficiency, International Tax and Public Finance, TD No. 732 (November 2009).

I. BUONO and G. LALANNE, The effect of the Uruguay Round on the intensive and extensive margins of trade, Journal of International Economics, TD No. 835 (February 2011).

G. BARONE, R. FELICI and M. PAGNINI, Switching costs in local credit markets, International Journal of Industrial Organization, TD No. 760 (June 2010).

E. COCOZZA and P. PISELLI, Testing for east-west contagion in the European banking sector during the financial crisis, in R. Matoušek; D. Stavárek (eds.), Financial Integration in the European Union, Taylor & Francis, TD No. 790 (February 2011).

S. NERI and T. ROPELE, Imperfect information, real-time data and monetary policy in the Euro area, The Economic Journal, TD No. 802 (March 2011).

M. AFFINITO, Do interbank customer relationships exist? And how did they function in the crisis? Learning from Italy, Journal of Banking and Finance, TD No. 826 (October 2011).

O. BLANCHARD and M. RIGGI, Why are the 2000s so different from the 1970s? A structural interpretation of changes in the macroeconomic effects of oil prices, Journal of the European Economic Association, TD No. 835 (November 2011).

R. CRISTADORO and D. MARCONI, Households Savings in China, Chinese Economic and Business Studies, TD No. 838 (November 2011).